Exam-integrity challenges are increasing worldwide, particularly in developing countries where resources for secure assessment are scarce. This research proposes an ethical, AI-based proctoring system designed to enhance online exam fairness through real-time prevention and detection and propose a real‑time, AI‑driven proctoring system built with React for the front‑end, Python for the back-end, MySQL for secure data storage, and computer‑vision plus machine‑learning models for behavior detection. Prior to the exam, candidates register and verify their identity via One Time Passcode and face-recognition. During the test, the examinee’s webcam feed is processed by lightweight convolutional neural networks that recognize common dishonest actions such as looking away from the screen, using prohibited objects. Simultaneously, the microphone monitors for spoken responses, while client-side controls block copy-paste, detect tab-switching and flag multi-application usage. When the system flags a suspicious event, it instantly sends a brief, non‑intrusive alert and stores the relevant image as evidence for a human proctor to review and decide whether further action is needed. By integrating automated monitoring with human oversight, the system offers a scalable, cost‑effective solution for universities, language schools, and any organization that conducts online assessments, including essay‑writing tests. Expected outcomes include higher exam fairness, improved trust among students and institutions, and a practical model that can be deployed in resource‑limited settings supporting displaced learners worldwide. This study advances the field of AI‑assisted proctoring and contributes to equitable online examination practices.

Authors:
Ohnmar Nyunt, Kobe Institute of Computing, Japan

About the Presenter(s)
Ohnmar Nyunt, a Master’s student in the ICT Innovator Course at Kobe Institute of Computing, Japan, focuses on ICT innovation, education, and lifelong learning. Her current project enhances online exam integrity through an AI-based proctoring system.

Connect on Linkedin
https://www.linkedin.com/in/ohnmar-nyunt/

See this presentation on the full schedule – Wednesday Schedule

Bookmark